Introduction
Current sepsis diagnostic criteria lose precision at high altitude due to physiological variation in vital functions. Because of that, it is necessary to elaborate new sepsis diagnostic criteria, especially designed for its use at populations living at 2000 meters above sea level or higher altitudes.
Methods
This is a prospective, longitudinal and observational study, that took place at EsSalud Regional Hospital at Huancayo. Two cohorts were studied: a derivation cohort (394 patients) and a validation cohort (123 patients). A total of 147 adult patients with sepsis diagnosis and 247 healthy controls were included in the derivation cohort. Sepsis was confirmed with cultures, and diagnosis of infection was made according to CDC and McGeer criteria, in patients admitted to Intensive and Intermediate Care Unit from January 2007 to April 2008. Each case was followed up to discharge or death. With the obtained data, new diagnostic models were devised. A validation cohort of 72 patients with sepsis, 51 hospitalized patients without sepsis and 120 nonhospitalized normal subjects was developed (validation cohort) with patients admitted to Internal Medicina Seervice from January 2012 to December 2013. Diagnostic criteria were evaluated with 2x2 tables, and specificity, sensitivity, negative and positive predictive values, as well as ROC AUC values were determined. Relative risk values and 95% confidence intervals are presented.
Results
A total of 10 physiological variables were obtained to develop the new criteria. Three of them were selected: respiratory rate > 23, temperature > 37.8 °C or 37.5 °C, and heart rate > 85 beats/minute. More specific alternative diagnostic criteria selected were: Criterium 10 (ALT10) characterized by a RR > 23, HR > 85, T > 37.5°C and Leukocytes >10000 or < 4000 or bands > 5% or neutrophils > o = 70%; Criterion 22 (ALT 22): RR >23, HR > 85, T > 37.8°C, Leukocytes > 10000 o < 4000 or bands > 5% or neutrophils > o = 70%. ALT10 and ALT 22 sensitivity were superior to that of classical criteria: 100% and 99.3% versus 95.1%, respectively. Validation in hospitalized patients without sepsis revealed similar results, with ALT10 and ALT22 reaching 100% sensitivity. Specificity remained around 83.7 to 84.8% when using the two alternative criteria in the derivation cohorts, while it was 72% when using classical criteria.
Conclusions
Current sepsis criteria have limited specificity for diagnosis at high – altitude populations. The use of the new diagnostic criteria may decrease the rate of nondiagnosed patients from 5 to 0%, while increasing specificity from 72 to 84.8%.